Organometallics 2005, 24, 1173-1183 1173

Synthesis, Structure, and Ion Pair Dynamics of

β-Diketiminato-Supported Organoscandium Contact Ion
Pairs
Paul G. Hayes, Warren E. Piers,* and Masood Parvez
Department of Chemistry, University of Calgary, 2500 University Drive NW,
Calgary, Alberta, T2N 1N4, Canada

Received January 4, 2005

A rare family of base-free organoscandium alkyl cations stabilized by β-diketiminato ligands

(Ar)NC(R)CHC(R)N(Ar) (Ar ) 2,6-iPr-C6H3; R ) CH3 (LMe), R ) tBu (LtBu)) has been prepared
by reaction of LScR2 with perfluorinated boranes B(C6F5)3 and (C12F8)B(C6F5). While the
LMe ancillary lacked sufficient steric bulk to prevent C6F5 transfer to the metal center, the
LtBu derivatives were quite robust in solution. Although these species were subject to thermal
decomposition via metalation of an isopropyl group and loss of RH (∆Sq ) -7.0(7) eu and
∆Ηq ) 21.5(2) kcal mol-1), the process was not sufficiently rapid to hinder the development
of their rich organometallic chemistry. The solution and solid state structures were studied
in detail, revealing the presence of both exo- and endo-MeB(ArF)3 isomers if the metal alkyl
was small. Detailed NMR studies (low-temperature EXSY) allowed for observation of both
inter- and intramolecular ion pair reorganization processes, whose mechanism is discussed
in terms of the involvement of “solvent separated ion pairs”. This is the first such study
involving non-metallocenium group 3 metal based ion pairs.

Introduction activation parameters associated with the process(es)

Organometallic cations of the early transition metals,
particularly those of group 4, have been studied inten-
sively over the past couple of decades due to their
importance as catalytic species in a variety of reactions,
including olefin polymerization,1 C-H bond function-
alization,2 and Si-H dehydropolymerization.3 Generally
these cations are formed via reaction of neutral organ-
ometallic precursors with strong Lewis acids such as
perfluoroarylboranes,4 or trityl salts of weakly coordi-
nating anions.5 The static structural features of many
of these ion pairs are well established,2,6 but debate
concerning the dynamic behavior of these species in
solution, processes that have important consequences
for catalyst behavior in the polymerization reactions
mentioned above,7 has been lively.
Trends in the rates of exchange of diastereotopic
groups on the supporting ligand framework and the
* To whom correspondence should be addressed. E-mail: wpiers@
ucalgary.ca.
(1) Chen, E. Y. X.; Marks, T. J. Chem. Rev. 2000, 100, 1391.
(2) (a) Jordan, R. F.; Taylor, D. F. J. Am. Chem. Soc. 1989, 111,
778. (b) Rodewald, S.; Jordan, R. F. J. Am. Chem. Soc. 1994, 116, 4491.
(3) (a) Sadow, A. D.; Tilley, T. D. Organometallics 2001, 20, 4457.
(b) Sadow, A. D.; Tilley, T. D. Organometallics 2003, 22, 3577. (c)
Sadow, A. D.; Tilley, T. D. J. Am. Chem. Soc. 2003, 125, 9462.
(4) (a) Yang, X. M.; Stern, C. L.; Marks, T. J. J. Am. Chem. Soc.
1991, 113, 3623. (b) Ewen, J. A.; Elder, M. J. CA-A 1991, 2, 27. (c)
Ewen, J. A.; Elder, M. J. Chem. Abstr. 1991, 115, 136998g; 256895t.
(d) Hlatky, G. G.; Turner, H. W. In PCT Int. Appl. W/O 91/14713; Exxon
Chemical Co., 1991. (e) Piers, W. E. Adv. Organomet. Chem. 2005, 52,
1.
(5) Chien, J. C. W.; Rausch, M. D.; Tsai, W.-M. J. Am. Chem. Soc.
1991, 113, 8570.
(6) Yang, X. M.; Stern, C. L.; Marks, T. J. J. Am. Chem. Soc. 1994,
116, 10015.
(7) See: Chen, M.-C.; Roberts, J. A. S.; Marks, T. J. J. Am. Chem.
Soc. 2004, 126, 4605, and references therein.
involved have been used to infer mechanistic informa-
tion concerning the nature of the dynamic behavior
associated with group 4 metallocenium ion pairs. For
example, Marks et al. have developed an elegant
1,1′,2,2′-tetramethylzirconocene probe to characterize
two dynamic processes within ion pairs incorporating
the [H3CB(C6F5)3]- anion.6 The first exchanges both the
diastereotopic ring methyl groups and the abstracted
B-CH3/nonabstracted Zr-CH3 pair via “B(C6F5)3 dis-
sociation/reabstraction” (Scheme 1). The second process,
referred to as “ion pair symmetrization”, involves dis-
sociation of the entire methylborate anion from the
contact ion pair (CIP), possibly into a solvent-separated
ion pair (SSIP), followed by recombination on the other
side of the zirconocene wedge. This process exchanges
only the diastereotopic ring methyl groups and not the
Zr-Me/B-Me pair.
Similar exchange phenomena have been examined in
metallocenium ions with other ligand sets featuring
diastereotopic groups8-10 such that a range of data
concerning these processes is available. The intimate
details of the mechanisms of the two processes are
profoundly influenced by variables such as metallo-
cenium ion structure, solvent/medium polarity and ionic
strength, the concentration regime employed, and the
stoichiometry of the ion pair generating reaction.11
Furthermore, the role of ion pair aggregates in this
(8) (a) Beck, S.; Geyer, A.; Brintzinger, H.-H. Chem. Commun. 1999,
2477. (b) Beck, S.; Lieber, S.; Schaper, F.; Geyer, A.; Brintzinger, H.-
H. J. Am. Chem. Soc. 2001, 123, 1483. (c) Schaper, F.; Geyer, A.;
Brintzinger, H. H. Organometallics 2002, 21, 473.
(9) Mohammed, M.; Nele, M.; Al-Humydi, A.; Xin, S.; Stapleton, R.
A.; Collins, S. J. Am. Chem. Soc. 2003, 125, 7930.
(10) Song, F.; Cannon, R. D.; Bochmann, M. J. Am. Chem. Soc. 2003,
125, 7641.

10.1021/om050007v CCC: $30.25 © 2005 American Chemical Society

Publication on Web 02/16/2005
1174 Organometallics, Vol. 24, No. 6, 2005
Scheme 1
exchange process has been debated and probed by a
variety of sophisticated NMR techniques, including
EXSY (exchange spectroscopy) and 1D DPFGSE-NOE
(differential pulse field gradient spin-echo nuclear
Overhauser effect) experiments.12 The picture that
accrues from these studies is that in the less polarized
CIPs, or “inner sphere ion pairs”,12b no evidence for ion
pair aggregation is found even at relatively high con-
centrations, while for more charge-separated SSIPs
(“outer sphere ion pairs”) aggregation into ion pair
quadruples (or multiples) is substantial at concentra-
tions as low as 1-2 mM, which is substantially higher
than the concentrations typically employed in industri-
ally relevant olefin polymerization processes. Neverthe-
less, such aggregates might be expected to play a role
in the observed ion pair reorganization processes in
metallocenium ions, and while explicit evidence for the
role of such aggregates in ion pair reorganization
processes has been lacking, it is clear from the above
studies12b that significantly greater ion mobility is to
be expected in the SSIPs than the CIPs, in which fairly
strong anion-cation interactions are featured.
In contrast to the extensive developments concerning
the group 4 metal organometallic cations, fewer ex-
amples of cations based on the group 3 or lanthanide
metals have been presented, and even less is known
about their dynamic behavior.13 Metallocenium ions of
the group 3 metals of general formula [Cp2MIII]+[A]-
have been prepared14 and function as models for weak
cation-anion or cation-solvent interactions, but of
course do not retain a reactive alkyl group. Thus, to
develop the chemistry of well-defined alkyl cations of
group 3 metals, a monoanionic supporting ligand of
sufficient steric bulk is required; examples of suitable
ligands have only recently emerged, allowing for de-
tailed study of this family of compounds,15 but even in
(11) (a) Deck, P. A.; Marks, T. J. J. Am. Chem. Soc. 1995, 117, 6128.
(b) Deck, P. A.; Beswick, C. L.; Marks, T. J. J. Am. Chem. Soc. 1998,
120, 1772. (c) Beswick, C. L.; Marks, T. J. Organometallics 1999, 18,
2410. (d) Beswick, C. L.; Marks, T. J. J. Am. Chem. Soc. 2000, 122,
10358.
(12) (a) Stahl, N. G.; Zuccaccia, C.; Jensen, T. R.; Marks, T. J. J.
Am. Chem. Soc. 2003, 125, 5256. (b) Zuccaccia, C.; Stahl, N. G.;
Macchiono, A.; Chen, M.-C.; Roberts, J. A.; Marks, T. J. J. Am. Chem.
Soc. 2004, 126, 1448.
(13) Piers, W. E.; Emslie, D. J. H. Coord. Chem. Rev. 2002, 233-
234, 131.
(14) (a) Hazin, P. N.; Bruno, J. W.; Schulte, G. K. Organometallics
1990, 9, 410. (b) Bouwkamp, M. W. Ph.D. Thesis, University of
Groningen, 2004.
Hayes et al.
these instances, few are free of coordinating bases.
Interest in such cations stems from the presumption
that they would exhibit the same boost in reactivity that
group 4 metal-based cations display in comparison to
their neutral counterparts. Several recent publications
have proven this to be the case; where direct comparison
is possible, cationic group 3 compounds show both
increased activity and novel selectivity in, for example,
scandium-catalyzed ethylene/olefin copolymerizations,15a
olefin and alkyne hydroaminations,16 or alkyne dimeriza-
tions.15b
Our group has exploited the bulky β-diketiminato
ligand framework, an ancillary ligand that has gained
considerable importance over the past decade,17 to
stabilize a variety of base-free dialkyl organoscandium
compounds.18,19 The ligand incorporating methyl groups
in the diketimine backbone is slightly less sterically
bulky than the tBu-substituted version, but both can
support formally eight-electron dialkyl complexes that
adopt the out-of-plane structure shown in I. This
engenders diastereotopicity in the Sc-R groups, as well
as the N-Ar substituents in the slow exchange limiting
structure, but exchange of the diastereotopic groups is
rapid on the NMR time scale. Furthermore, in the CIPs
(15) (a) Bambirra, S.; Bouwkamp, M. W.; Meetsma, A.; Hessen, B.
J. Am. Chem. Soc. 2004, 126, 9182. (b) Tazelaar, C. G. J.; Bambirra,
S.; van Leusen, D.; Meetsma, A.; Hessen, B.; Teuben, J. H. Organo-
metallics 2004, 23, 936. (c) Bambirra, S.; van Leusen, D.; Meetsma,
A.; Hessen, B.; Teuben, J. H. Chem. Commun. 2001, 637. (d) Arndt,
S.; Zeimentz, P. M.; Spaniol, T. P.; Okuda, J.; Honda, M.; Tatsumi, K.
Dalton Trans. 2003, 3622. (e) Arndt, S.; Spaniol, T. P.; Okuda, J.
Organometallics 2003, 22, 775. (f) Arndt, S.; Spaniol, T. P.; Okuda, J.
Chem. Commun. 2002, 896. (g) Cameron, T. M.; Gordon, J. C.;
Michalczyk, R.; Scott, B. L. Chem. Commun. 2003, 2282. (h) Schav-
erien, C. J. Organometallics 1992, 11, 3476. (i) Hajela, S.; Schaefer,
W. P.; Bercaw, J. E. J. Organomet. Chem. 1997, 532, 45. (j) Hayes, P.
G.; Welch, G. C.; Emslie, D. J. H.; Noack, C. L.; Piers, W. E.; Parvez,
M. Organometallics 2003, 22, 1577. (k) Lawrence, S. C.; Ward, B. D.;
Dubberley, S. R.; Kozak, C. M.; Mountford, P. Chem. Commun. 2003,
2880.
(16) Lauterwasser, F.; Hayes, P. G.; Piers, W. E.; Brase, S.; Schafer,
L. L. Organometallics 2004, 23, 2234.
Ion Pair Dynamics Organometallics, Vol. 24, No. 6, 2005 1175

Scheme 2

generated from these neutral dialkyls and B(C6F5)3,20
which are highly active olefin polymerization catalysts,
this exchange phenomenon can be used to probe the ion
pair dynamic processes associated with these non-
metallocene ion pairs. Herein we fully describe the
synthesis and properties of these CIPs along with their
dynamic behavior in toluene solution.
Results and Discussion
Reactions of [LMeScMe2]2 and LtBuScMe2 with
B(C6F5)3. Reaction of dimeric [LMeScMe2]221 with 1
equiv of B(C6F5)3 in toluene at -35 °C afforded the
expected ion pair [LMeScMe]+[MeB(C6F5)3]- as moni-
tored by 1H, 11B, and 19F NMR spectroscopy. A broad
resonance at -16.8 ppm in the 11B NMR spectrum, in
combination with a ∆δm,p ) 4.2 ppm in the 19F NMR
spectrum, supported the formation of a loosely bound
CIP with a µ-methyl contact between the boron and
scandium centers.22 While the ion pair was stable at
-35 °C, upon gradual warming to room temperature,
-C6F5 transfer from the borate to scandium generated
the ion pair [LMeSc(C6F5)]+[Me2B(C6F5)2]- (Scheme 2)
presumably via the neutral complex LMeSc(Me)(C6F5)
and MeB(C6F5)2.23 This ion pair was highly soluble in
toluene and thermally unstable at ambient temperature
and so was characterized only in solution. However, a
more robust ion pair incorporating [MeB(C6F5)3]- in-
stead of the [Me2B(C6F5)2]- anion was isolated when 2
equiv of B(C6F5)3 was employed; here, the initial product
was the dicationic species [LMeSc]2+[MeB(C6F5)3]-2, as
evidenced by the disappearance of both Sc-CH3 reso-
nances in the 1H NMR spectrum and the emergence of
two sets of [MeB(C6F5)3]- signals in the 19F NMR
spectrum. This doubly activated species24 is also prone
to C6F5 transfer, ultimately giving rise to endo-[LMeSc-
(C6F5)]+[MeB(C6F5)3]- (Scheme 2),25 which was charac-
terized by X-ray crystallography (Figure 1).
The solid state structure of endo-[LMeSc(C6F5)]+[MeB-
(C6F5)3]- exhibits a symmetrically bound β-diketiminato
ancillary with the metal lying 0.645(5) Å from the plane
formed by the C3N2 ligand backbone. An elongated Sc-
C1 distance of 2.699(4) Å (cf. an average Sc-C distance
of 2.29 Å in [LMeScMe2]2) is indicative of a weakly
coordinating [MeB(C6F5)3]- anion. The nonlinear Sc-
C1-B angle of 133.42(25)° is atypical of metallocenium
ion pairs with this counteranion,1 but this bending
allows for an interaction between Sc and F1, as evi-
denced by a short contact of 2.221(2) Å.26 This dative
distance is comparable to those observed in [LtBuScMe]+-
[MeB(C6F5)3]- (2.390(4)Å)20b and[Cp*2Sc(FC6H5)2]+[BPh4]-
(2.2725(17), 2.2884(16) Å),14b but longer than the bridg-
ing Sc-F distances (2.039(23)-2.070(24) Å) found in
[Cp2ScF]3.27 The C33-F1 bond distance of 1.413(5) Å is
concomitantly elongated relative to the other noncoor-
dinating C-F bonds, which average to 1.349(5) Å; the
remaining Sc-C and Sc-N distances agree well with
typical values.
The low-temperature 1H and 19F NMR spectra of
[L Sc(C6F5)]+[MeB(C6F5)3]- do not indicate a solution
Me
structure consistent with that observed by X-ray crys-

(17) Bourget-Merle, L.; Lappert, M. F.; Severn, J. R. Chem. Rev.

2002, 102, 3031.
(18) (a) Hayes, P. G.; Piers, W. E.; Lee, L. W. M.; Knight, L. K.;
Parvez, M.; Elsegood, M. R. J.; Clegg, W. Organometallics 2001, 20,
2533. (b) Knight, L. K.; Piers, W. E.; McDonald, R. Chem. Eur. J. 2000,
6, 4322. (c) Knight, L. K.; Piers, W. E.; Fleurat-Lessard, P.; McDonald,
R.; Parvez, M. Organometallics 2004, 23, 2087.
(19) Other scandium β-diketiminato complexes: (a) Basuli, F.;
Tomaszewski, J.; Huffman, J. C.; Mindiola, D. J. Organometallics 2003,
22, 4705. (b) Neculai, A. M.; Neculai, D.; Roesky, H. W.; Magull, J.;
Baldus, M.; Andronesi, O.; Jansen, M. Organometallics 2002, 21, 2590.
(20) (a) Lee, L. W. M.; Piers, W. E.; Elsegood, M. R. J.; Clegg, W.;
Parvez, M. Organometallics 1999, 18, 2947. (b) Hayes, P. G.; Piers,
W. E.; McDonald, R. J. Am. Chem. Soc. 2002, 124, 2132.
(21) Hayes, P. G.; Piers, W. E.; Parvez, M. J. Am. Chem. Soc. 2003,
125, 5622.
(22) Horton, A. D. Organometallics 1996, 15, 2675.
(23) Identified by 1H NMR spectroscopy: Spence R. E. v H.; Piers,
W. E.; Sun, Y.; Parvez, M.; MacGillivray, L. R.; Zaworotko, M. J.
Organometallics 1998, 17, 2459.
(24) Guérin, F.; Stephan, D. W. Angew. Chem., Int. Ed. 2000, 39,
1298.
(25) Because the scandium center deviates from the β-diketiminato
ligand plane in these complexes, the two remaining ligand coordination Figure 1. Crystalmaker depiction of the molecular struc-
sites are diastereotopic. We refer to the site pointing away from the
C3N2 framework as the exo position and that situated underneath the
ture of endo-[LMeSc(C6F5)]+[MeB(C6F5)3]-. Hydrogen atoms
C3N2 ring as the endo site. In the CIPs discussed herein, we refer to and all of the atoms of the front N-aryl group (except for
exo and endo isomers as those with the remaining alkyl group in this ipso-C(20)) are removed for clarity. Selected bond distances
position. (Å): Sc-N(1), 2.094(3); Sc-N(2), 2.087(3); Sc-C(1), 2.699(4);
(26) Song, X.; Thornton-Pett, M.; Bochmann, M. Organometallics
1998, 17, 1004.
Sc-C(50), 2.220(5); Sc-F(1), 2.221(2). Selected bond angles
(27) Bottomley, F.; Paez, D. E.; White, P. S. J. Organomet. Chem. (deg): N(1)-Sc-N(2), 87.9(3); Sc-C(1)-B, 133.4(2); C(1)-
1985, 291, 35. Sc-C(50), 111.7(2).
1176 Organometallics, Vol. 24, No. 6, 2005
Scheme 3
tallography; there are no upfield resonances28 pointing
to an Sc-F interaction in the low temperature (-100
°C) 19F NMR spectrum. Rather, two isomers are ob-
served in a 7:3 ratio, corresponding to the diastereomers
in which the -C6F5 group occupies the endo or exo
positions.29
Pentafluorophenyl group transfer from the [H3CB-
(C6F5)3]- anion to sterically accessible electrophilic
centers has been observed in several other cases.30 Since
this is essentially an undesirable deactivation pathway
in the context of olefin polymerization, we turned to the
more sterically encumbered scandium dialkyl cations
supported by the LtBu ligand in hopes of disfavoring this
reaction path. As reported previously,20b the CIPs
obtained via reaction of the monomeric dimethyl com-
pound LtBuScMe2, 1a, and 1 or 2 equiv of B(C6F5)3
(henceforth referred to as B1) are immune to -C6F5
transfer under ambient conditions, due to the greater
steric congestion provided by this ligand in comparison
to LMe, and form well-behaved ion pairs (Scheme 3).
Thus, the mono methyl ion pair 2a-B1 is formed as a
stable, isolable yellow solid, and its structure has been
determined by X-ray crystallography. As in the case of
the endo-[LMeSc(C6F5)]+[MeB(C6F5)3]- ion pair discussed
above, the isomer 2a-B1 that crystallizes is the endo
isomer,20b in which the bulky [H3CB(C6F5)3]- anion
occupies the exo site via a similar coordination mode
featuring a C-F f Sc interaction.20b Ion pair 2a-B1 is
stable in toluene for prolonged periods below 0 °C and
indefinitely as a solid under an inert atmosphere if
stored at -35 °C. Although no sign of decomposition via
C6F5 transfer occurred upon warming to room temper-
ature, 2a-B1 slowly evolved methane through a meta-
lation pathway involving a methyl C-H bond of one of
the N-aryl isopropyl groups, similar to that observed for
the neutral dialkyl species (vide infra).
(28) (a) Ruwwe, J.; Erker, G.; Fröhlich, R. Angew. Chem., Int. Ed.
Engl. 1996, 35, 80. (b) Temme, B.; Karl, J.; Erker, G. Chem. Eur. J.
1996, 2, 919. (c) Temme, B.; Erker, G.; Karl, J.; Luftmann, H.; Fröhlich,
R.; Kotila, S. Angew. Chem., Int. Ed. Engl. 1995, 34, 1755.
(29) The assignment of isomers was not rigorously established in
this case.
(30) (a) Korolev, A. V.; Ihara, E.; Guzei, I. A.; Young, V. G., Jr.;
Jordan, R. F. J. Am. Chem. Soc. 2001, 123, 8291. (b) Spitzmesser, S.
K.; Gibson, V. C. J. Organomet. Chem. 2003, 673, 95. (c) Walker, D.
A.; Woodman, T. J.; Hughes, D. L.; Bochmann, M. Organometallics
2001, 20, 3772. (d) Walker, D. A.; Woodman, T. J.; Schormann, M.;
Hughes, D. L.; Bochmann, M. Organometallics 2003, 22, 797.
Hayes et al.
Further activation of this cation is observed when a
second equivalent of B1 is added, forming the unusual
dication 3a-B1 as a pale yellow solid; treatment of this
species with 1 equiv of neutral dimethyl complex 1a
regenerates the monomethyl cation 2a-B1. Such “double
activation” reactions are rare, but have precedence in
group 4 metal systems with strongly donating ligands.24
The solution 1H NMR data for 3a-B1 were diagnostic,
with no Sc-Me resonances in evidence and two distinct
[MeB(C6F5)3]- signals at 1.55 and 1.92 ppm for the
chemically diastereotopic exo and endo anions. The 19F
NMR spectrum also exhibited two sets of C6F5 reso-
nances with ∆δm,p ) 4.3 and 5.8 ppm (assigned by 19F-
19F 2D COSY). These values correlate well with those
observed for the two isomers of 2a-B1. While a 19F-
19F 2D EXSY experiment confirmed the two anions are
exchanging, it was not possible to coalesce the process
in one-dimensional 1H NMR experiments.
As is also shown in Scheme 3, all of the above
chemistry can be repeated using the similarly Lewis
acidic perfluoro-9-borafluorene species (C12F8)B(C6F5),
B2.31 Analogous products 2a-B2 and the doubly acti-
vated dication 3a-B2 were formed and exhibited es-
sentially the same properties as the closely related B1
series. These ion pairs were prepared for use in the ion
pair dynamic studies described below.
Because the two substituents associated with the
scandium center in ion pairs 2a-Bn are different, these
compounds exist in solution as an exchanging pair of
diastereomers, which are indistinguishable by 1H or 19F
NMR spectroscopy at room temperature. However, as
samples of 2a-Bn are cooled, coalescence behavior is
observed in the 1H and 19F NMR spectra, and the
exchange is frozen out to reveal spectra consistent with
the presence of exo and endo isomers in a 2.2:1 ratio as
shown in Scheme 3 and the representative 1H NMR
spectra for 2a-B1 in Figure 2. The barrier to exchange
between the two diastereomers was estimated to be
12.4(5) kcal mol-1 based on analysis of the coalescence
behavior of the spectra, a barrier comparable to those
observed for exchange of endo and exo alkyl groups in
the neutral dialkyl compounds. As will be seen, the
observed exchange between endo-2a-B1and exo-2a-
(31) Chase, P. A.; Piers, W. E.; Patrick, B. O. J. Am. Chem. Soc.
2000, 122, 12911.
Ion Pair Dynamics
Figure 2. 1H NMR Spectra of endo/exo-2a-B1 at 263 K
(near coalescence, top) and 213 K (slow exchange regime,
bottom).
Figure 3. 1H-1H 2D ROESY NMR spectrum of endo/
exo-2b-B1.
B1 offers an opportunity to probe the ion pair dynamics
in these non-metallocene ion pairs.
As in the case of [LMeSc(C6F5)]+[MeB(C6F5)3]-, no
evidence for the [(µ-CH3)(µ-FC6F4)B(C6F5)2]- bonding
mode of the methylborate anion was observed in solu-
tion, suggesting that the normal µ-CH3 mode, in which
a more or less linear Sc-C-B arrangement occurs, is
predominant in solution. When the sample of Figure 2
was analyzed via a 1H-1H 2D-ROESY NMR experi-
ment,32 distinct cross-peaks were observed between one
of the major isomer isopropyl methine groups and the
B-CH3 group and a minor isomer isopropyl resonance
and the Sc-CH3 signal (Figure 3). On the basis of an
analysis of the solid state structures of several scandium
β-diketiminato derivatives, the Sc-Cendo carbon atom is
closest to the lower isopropyl methine carbons, as
illustrated in II. The Sc-Cexo carbon is invariably further
Organometallics, Vol. 24, No. 6, 2005 1177
Figure 4. Crystalmaker depiction of exo-2b-B1. Hydro-
gen atoms and all atoms of the front N-aryl group (except
for ipso-C(8)) are removed for clarity. Selected bond dis-
tances (Å): Sc-N(1), 2.072(2); Sc-N(2), 2.086(2); Sc-C(1),
2.499(2); Sc-C(2), 2.133(3). Selected bond angles (deg):
N(1)-Sc-N(2), 96.65(7); Sc-C(1)-B, 176.8(2); C(1)-Sc-
C(2), 113.5(1); Sc-C(2)-Si(1), 144.5(2).
removed from these isopropyl groups by ∼0.5 Å, while
the upper isopropyl methine carbons are even further
away at ∼4.8 Å. We thus tentatively assign the major
solution isomer as exo-2a-B1 and the minor diastere-
oisomer as the endo-2a-B1 even though the compound
crystallizes preferentially as the endo-2a-B1 struc-
ture.33 We attribute this observation to the fact that the
bonding mode of the anion in the solid is different from
that observed in solution.
The boranes B1 and B2 also react rapidly with the
bis(trimethylsilylmethyl) organoscandium complex 1b18a
to give ion pairs 2b-Bn, in which the two silyl alkyl
groups have coupled, and the counteranion is again the
methylborate species [H3CB(C6F5)3]-. This reaction path
was indicated by the pattern of three resonances upfield
of 0 ppm in a 9:6:2 ratio observed in the 1H NMR
spectrum of 2b-B1 in d8-toluene. In addition, a broad
peak, integrating for 3H, was observed at 1.51 ppm, the
region where the methylborate signals in the previously
generated ion pairs are found. Interestingly, and unlike
ion pairs 2a, the 2b ion pairs exist as only one diaste-
reoisomer at all temperatures, as determined by 1H, 11B,
and 19F NMR spectroscopy; furthermore, reaction with
a second equivalent of borane is not observed in this
case. An X-ray diffraction structure determination
showed this to be the exo isomer (Figure 4). Crystals of
this ion pair were grown from cold hexanes.
The solid state structure of exo-2b-B1 confirms the
presence of a bulky -CH2SiMe2CH2SiMe3 group in the
(32) The ROESY pulse sequence was more effective than the NOESY
experiment for these relatively high molecular weight ion pairs: Two-
Dimensional NMR Spectroscopy: Applications for Chemists and
Biochemists, 2nd ed.; Croasman, W. R., Carlson, R. M. K., Eds.; VCH
Publishing Inc.: New York, 1994; pp 329-341.
(33) It should be noted that in an earlier paper we assigned the
isomers of the unsymmetrically alkylated complex LtBuSc(CH3)CH2-
SiMe3 using the assumption that the exo position was that giving rise
to ROESY cross-peaks with the isopropyl methine signals. Thus, using
the present analysis, that original assignment was probably incorrect
and the major isomer in this neutral compound has the methyl group
in the endo position.
1178 Organometallics, Vol. 24, No. 6, 2005 Hayes et al.

Scheme 4

exo coordination site and exhibits several noteworthy
metrical parameters. The Sc-C1 bond distance of
2.499(3) Å is significantly elongated from the typical
2.20 Å observed in the neutral starting materials, but
is shorter than the analogous distance observed in either
[LMeSc(C6F5)]+[MeB(C6F5)3]- or endo-2a-B1. Because
the methylborate anion occupies the more sterically
hindered endo site, it assumes a more typical linear Sc-
C1-B contact (176.8(2)°) with the scandium cation. The
inability of the anion to flex and electronically stabilize
the metal center through an Sc-F interaction results
in a more Lewis acidic metal center, thus drawing the
methylborate anion closer. The steric impact of the
-CH2SiMe2CH2SiMe3 alkyl is apparent in the Sc-C2-
Si1 bond angle of 144.47(15)°, which is opened consider-
ably from the ideal tetrahedral value of 109.5°. That this
structure is maintained in solution is supported by a
1H-1H 2D-ROESY NMR experiment, which shows a
distinct cross-peak between the downfield isopropyl
methane resonance (the “lower” isopropyl, see II) and
the CH3 group of the methyl borate anion; no cross-
peaks between the isopropyl and alkyl groups are
observed.
Two reasonable mechanisms for the formation of ion
pairs exo-2b-Bn are depicted in Scheme 4. Path A has
been proposed to operate in similar couplings observed
in early metal chemistry bis(trimethylsilyl) compounds34
and involves borane abstraction of a methyl group from
a silicon center in the CH2SiMe3 group; as the positive
charge develops on Si, it is quenched by transfer of the
other Sc-R′ group to form the coupled alkyl ligand and
the methylborate counteranion directly. The other
invokes direct abstraction of the entire CH2SiMe3 from
scandium by the boranes, yielding an intermediate ion
pair which can undergo β-methyl elimination22,35 from
the remaining alkyl group. The eliminated isobutylene
or dimethylsilene then undergoes attack by the
[Me3SiCH2B(C6F5)3]- anion to generate the neutral
dialkyl shown, which reacts immediately with the Bn
to yield the products. Although this mechanism has
some precedence in the decomposition of some L2Pt(CH2-
SiMe3)2 compounds,36 and borane abstraction of entire
CH2SiMe3 groups has been observed in some in-
stances,37 we currently favor path A. This mechanism
more readily accommodates the observations that the
bis-neopentyl complex 1c does not react at all with the
boranes B1 or B238 and that reaction of the mixed alkyl
species 1d with B1 exclusively produces the ion pair
exo-2d-B1, in which the neopentyl group migrates to
the silicon atom. Furthermore, for path B, observation
of intermediate ion pairs and small amounts of isobu-
tylene and/or H2CdSiMe2 decomposition products might
be expected; none of these species were observed in
these reactions.
As mentioned above, -C6F5 transfer from the anion
to the scandium center is not a feature of the ion pairs
formed when using LtBu. The greater steric presence of
this ligand is enough to discourage this decomposition
mode; however, as in the neutral dialkyl systems, these
compounds are prone to metallative loss of RH to form
an endo/exo mixture of the compounds 4-Bn (Scheme
5). The progress of these reactions was conveniently
followed using 19F and 1H NMR spectroscopy and were
found to proceed first order in the concentration of ion
pair.39 The observed rate constants for metalation in ion
pairs 2a-Bn and exo-2b-Bn at 308 K are collected in
Table 1. Representative rate constants for metalation
Scheme 5

(34) (a) Gielens, E.; Tiesnitsch, J. Y.; Hessen, B.; Teuben, J. H.

Organometallics 1998, 17, 1652. (b) Cameron, T. M.; Gordon, J. C.;
Michalcyzk, R.; Scott, B. L. Chem. Commun. 2003, 2282.
(35) Hajela, S.; Bercaw, J. E. Organometallics 1994, 13, 1147.
(36) Ankianiec, B. C.; Christou, V.; Hardy, D. T.; Thomson, S. K.;
Young, G. B. J. Am. Chem. Soc. 1994, 117, 9963.
(37) Arndt, S.; Spaniol, T. P.; Okuda, J. Chem. Commun. 2002, 896.
(38) Abstraction of a methyl group from a carbon atom by a
perfluoroaryl borane has no precedent to our knowledge and is probably
kinetically and thermodynamically disfavored.
(39) Representative 19F NMR spectra, first-order plots, and the
Eyring plot are included as Supporting Information.
Ion Pair Dynamics Organometallics, Vol. 24, No. 6, 2005 1179

Scheme 6

Table 1. Half-Lives and kexp for Metalation Dynamic Processes in Ion Pairs 2a/b-Bn. The
Reactions existence of exchanging exo and endo isomers in the
cmpd temp (K) t1/2 (h) kexp (s-1) methyl cations 2a-Bn offers an opportunity to explore
Ion Pairs the mechanism(s) of exchange via dynamic NMR spec-
2a-B1 308 1.83 1.05(2) × 10-4 troscopic techniques. In the present system, possible
2a-B2 308 1.32 1.46(2) × 10-4 exchange mechanisms include the borane dissociation/
2b-B1 308 1.97 9.77(3) × 10-5 reabstraction and ion pair reorganization mechanisms
2b-B2 308 7.70 2.50(3) × 10-5
2b-B1 308 2.0 1.02(3) × 10-4 characterized by Marks and others for the metalloce-
2b-B1 308 2.0 9.77(3) × 10-5 nium ion pairs (Scheme 1) or a simple “ligand flip”
2b-B1 313 1.2 1.67(2) × 10-4 mechanism similar to what is observed in the neutral
2b-B1 318 0.6 3.00(2) × 10-4 dialkyl precursors (Scheme 6). We thus probed the
2b-B1 323 0.4 5.53(3) × 10-4
dynamic behavior of these CIPs, using d8-toluene as a
Neutral Compounds medium.
1a 308 12.26 1.57(3) × 10-5
1b 308 17.99 1.07(2) × 10-5
The complete absence of borane dissociation/reab-
straction (which exchanges B-Me with Sc-Me groups)
in exo/endo-2a-B1 on any time scale was established
in the neutral analogues 1a and 1b are also included
by the following observations. A low-temperature 1H-
for comparison. The metalation of exo-2b-B1 was 1H 2D EXSY experiment exhibited exchange only be-
followed at various temperatures, and an Eyring plot39
tween the two Sc-Me signals, with no cross-peaks
yielded the activation parameters ∆Ηq ) 21.5(2) kcal
observed between the Sc-Me and B-Me resonances.
mol-1 and ∆Sq ) -7.0(7) eu. The enthalpic barrier is
However, B-Me/B-Me cross-peaks were also absent in
similar to that observed for the neutral analogues (cf.
this experiment, probably due to quadrupolar broaden-
19.7(6) kcal mol-1), but the entropy of activation, while
ing from the boron atoms, and so to unambiguously
still negative, is significantly lower in magnitude (cf.
ascertain if this process was taking place on the chemi-
-17(2) eu). This can likely be attributed to the enhanced
cal time scale, the deuterium labeling experiment
spatial freedom of the weakly coordinating [MeB(C6F5)3]-
depicted in Scheme 7 was conducted. The dicationic
anion in the σ-bond metathesis transition state.
compound 3a-B1 was treated with d6-1a, labeled
As can be seen from the data in Table 1, the measured selectively in the Sc-Me positions, producing 2 equiv of
rates of metalation for the various ion pairs were d3-exo/endo-2a-B1. While the kinetic product of this
approximately an order of magnitude greater than those reaction would be expected to incorporate the deuterium
determined for the neutral dialkyl species, with half- label only in the Sc-Me position, if borane dissociation/
lives of 1-2 h at 35 °C. Although the rate of metalation reabstraction were occurring, the label should wash into
in the methyl cations incorporating B1 and B2 were the B-Me sites, while unlabeled Me groups should
similar, curiously, exo-2b-B2 was significantly more
resistant to metalation that the B(C6F5)3 derivative Scheme 7
exo-2b-B1. In any case, the rates of this process are
generally quite slow, in keeping with the general
observation that σ-bond metathesis reactions40 (even
intramolecular ones) are not prominent in non-Cp*
scandium complexes. For example, we have never
observed C-D bond activation of solvent molecules, and
even reactions of the LRScR2 compounds with H2 are
slow.41 This has been attributed to more ionic character
in the L-Sc bonding, giving rise to retracted d-orbitals
that are less available for interaction with C-H
σ-bonds.42 This is perhaps fortunate, in that the ion
pairs 2 are stable enough toward metalation that
detailed studies concerning the ion pair dynamics as-
sociated with these compounds could be performed.

(40) (a) Thompson. M. E.; Baxter, S. M.; Bulls, A. R.; Burger, B. J.;
Nolan, M. C.; Santarsiero, B. D.; Schaefer, W. P.; Bercaw, J. E. J. Am.
Chem. Soc. 1987, 109, 203. (b) Sadow, A. D.; Tilley, T. D. J. Am. Chem.
Soc. 2003, 125, 7971.
(41) Hayes, P. G. Ph.D. Thesis, University of Calgary, 2004.
(42) Duchateau, R.; van Wee, C. T.; Teuben, J. H. Organometallics
1996, 15, 2291.
1180 Organometallics, Vol. 24, No. 6, 2005 Hayes et al.

Figure 5. Top: 400 MHz 1H NMR spectrum of exo/

endo-2a-B1 at 213 K for comparison to Bottom: 400 Hz 1H
NMR spectrum of selectively labeled exo/endo-d3-2a-B1 Figure 6. 400 MHz 1H-1H EXSY NMR spectrum for a
at 213 K after 6 h, with periodic (hourly) warming to room 1:1 mixture of exo/endo-2a-B1 and exo/endo-2a-B2 at
temperature for 5-10 min. No exchange of CH3 groups into 213 K, focusing on the region where the CH proton of the
the Sc-Me positions is observed. β-diketiminato ligand resonates. The peak marked with (*)
is likely due to a small portion of µ-methyl dimer (see ref
materialize in the Sc-Me groups. As can be seen in Fig- 20b) since samples were generated with a slight deficit of
ure 5, signals for the Sc-Me groups in exo/endo-2a-B1 B(C6F5)3; it constitutes less than 5% of the total [Sc] and
is not involved in the exchanges observed by this experi-
are nearly completely absent43 in this sample, even after
ment.
6 h, wherein the sample had been warmed to RT
periodically. Clearly then, this process does not occur Scheme 8
in these ion pairs even on the chemical time scale. Since
the facility of borane dissociation/reabstraction has been
related to the relative magnitudes of the M-Me and
B-Me bond strengths,11 this observation implies that
the Sc-CH3 bonds in 1a are somewhat weaker than
those in the zirconocenium and hafnocenium ions where
this process is observable, a notion further supported
by the observed ease with which a second Sc-Me group
is abstracted from these cations.
Although these experiments determined it should be
possible to study ion pair symmetrization in the absence
of borane dissociation/reabstraction, the 2a system is
complicated by the fact that it is possible to exchange
the exo and endo isomers via a ligand flip process
(Scheme 6) as well as an ion pair reorganization
mechanism. In an attempt to deconvolute these pro-
cesses, ion exchange between the exo/endo ion pairs
differentiated by the borane B employed was studied.
Thus, a 1:1 mixture of exo/endo-2a-B1 and exo/ four of the isomers/compounds present in this mixture
endo-2a-B2 in d8-toluene was cooled to a temperature at low temperature (213 K). The EXSY map in Figure
regime where exo/endo exchange was slow on the NMR 6 also shows that all four isomers are exchanging with
time scale and studied via 1H-1H EXSY spectroscopy.44 each other under these conditions. At the same mixing
By observing exchange within each ion pair versus time, however, the cross-peaks for exchange within the
exchange between each ion pair, we hoped to differenti- ion pairs are more intense than those that arise from
ate between the ligand flip (intramolecular) and the ion exchange between the ion pairs, suggesting qualitatively
pair reorganization process (potentially intermolecular). that the ligand flip process is faster than intermolecular
The experiment is depicted in Scheme 8, and it is ion pair reorganization.
assumed that the properties described above for exo/ Rate constants for the various exchanges manifested
endo-2a-B1 also apply to the B2 ion pair. As can be in the EXSY spectrum of Figure 6 can be obtained via
seen in Figure 6, the signals for the backbone proton of analysis of the volumes of the cross-peaks as a function
the β-diketiminato ligand are baseline resolved for all of mixing time using an equation that accounts for the
fact that the exo and endo isomers are present in about
(43) Close inspection reveals that a small amount (∼1-2%) of a 2:1 ratio.44 The EXSY experiments can be performed
Sc-CH3 is present at the beginning of the experiment; the amount does at various temperatures allowing for analysis of the
not change over time, and we attribute this to a small amount of
unlabeled material present in the d6-2a-B1 starting material. thermodynamic parameters associated with each ex-
(44) Perrin, C. L.; Dwyer, T. J. Chem. Rev. 1990, 90, 935. change. These data are collected in Table 2, while a
Ion Pair Dynamics Organometallics, Vol. 24, No. 6, 2005 1181

Table 2. Kinetic and Thermodynamic Data for Exchange between exo/endo-2a-B1 and exo/endo-2a-B2
(intermolecular, entries 1-4; intramolecular, entries 5-6) and exo-2b-B1 and exo-2b-B2 (entry 7)
T kexp
entry exchange partners (K) (s-1) ∆HqHardReturn(kcal mol-1) ∆SqHardReturn(eu)
1 exo-2a-B1 exo-2a-B2 222.7 0.52(1) 10.5(3) -12(1)
228.0 0.73(2)
233.4 1.68(2)
238.7 2.50(2)
2 exo-2a-B1 endo-2a-B2 222.7 0.51(1) 11.6(3) -7(1)
228.0 0.80(1)
233.4 1.67(1)
238.7 3.04(3)
3 endo-2a-B1 exo-2a-B2 217.3 0.23(1) 11.8(2) -7(1)
222.7 0.50(1)
228.0 0.73(1)
233.4 1.59(3)
238.7 3.00(3)
4 endo-2a-B1 endo-2a-B2 217.3 0.37(1) 11.0(2) -9(1)
222.7 0.78(2)
228.0 1.27(1)
233.4 2.40(3)
238.7 4.19(4)
5 endo-2a-B1 exo-2a-B1 217.3 0.68(2) 13.2(2) 2(1)
222.7 2.30(3)
228.0 4.03(4)
233.4 7.12(6)
238.7 12.60(9)
244.0 26.9(2)
6 exo-2a-B2 endo-2a-B2 217.3 0.63(1) 13.2(3) 2(1)
222.7 2.02(2)
228.0 3.18(3)
233.4 5.82(6)
238.7 9.95(9)
244.0 26.7(2)
7 exo-2b-B1 exo-2b-B2 250.4 9.56(8) 8.7(3) -20(2)
244.0 5.20(4)
244.0 5.17(4)
244.0 5.12(4)
244.0 5.24(4)
237.0 2.57(3)
226.4 1.36(1)
221.3 0.85(1)
217.1 0.50(1)

representative Eyring plot (for the data concerning
exchange between endo-2a-B1 and endo-2a-B2, entry
4) is given in Figure 7.
As can be seen, the ∆Sq for the exchange processes
between the ions are negative in all cases, with the
magnitude varying from -7(1) to -12(1) eu at 0.0197
M. On the other hand, the ∆Sq values for intramolecular
exchange between isomers with the same anion (exo/
endo-2a-B1 and exo/endo-2a-B2, entries 5 and 6,
Table 2) are close to zero, with values similar to that
Figure 7. Eyring plot of intermolecular exchange between
endo-2a-B1 and endo-2a-B2 (Table 2, entry 4).
observed for the ligand flip process in the neutral dialkyl
derivative LtBuSc(CH2SiMe3)2, 1b (∆Ηq ) 12.9(2) kcal
mol-1, ∆Sq ) -1.6(5) eu). This observation, along with
4-fold rate enhancements for these exchanges relative
to the intramolecular examples, suggests that the ligand
flip process is the dominating exchange mechanism in
these situations. The ∆Hq are similar for all of the
exchange pathways with values ranging from 10.5 to
11.8 kcal mol-1, suggesting that the differences in the
rates are entropically driven.
The negative entropies of activation for the exchanges
between isomers of 2a-B1 and 2a-B2 are suggestive
of a mechanism with a rate-limiting step that is as-
sociative in character, although the values are generally
fairly low. Attempts to ascertain the effect of varying
[Sc] on the rates of exchange were unsuccessful due to
the narrow range of concentrations accessible under
these conditions and a threshold concentration (≈1.5 ×
10-2 M) necessary to obtain reliable rate data from an
EXSY experiment over a manageable time frame. Since
the ∆Sq values could be influenced by the ligand flip
process found in these scandium methyl cations, we
decided to explore the intermolecular exchange process
in ion pairs 2b-Bn and 2d-Bn, in which only the exo
isomer is observed and presumably intramolecular
ligand flip processes are suppressed relative to inter-
molecular ion exchange. That ion promiscuity is also a
1182 Organometallics, Vol. 24, No. 6, 2005
Figure 8. 1H-1H EXSY NMR spectrum of exo-2b-B1 and
exo-2b-B2.
characteristic of these ion pairs was demonstrated
qualitatively by mixing exo-2b-B2 and exo-2d-B1 in
a 1:1 ratio and recording the 1H NMR spectrum, which
showed a statistical mixture of all four possible ion pairs
(i.e., exo-2b-B1 and exo-2d-B2 were present as well)
upon mixing. This was demonstrated by computation-
ally summing the 1H NMR spectra of the four individual
ion pairs and comparing the result to the spectrum of
the mixture; there was excellent agreement between the
two spectra (see Figure S3).
This qualitative experiment was followed up with a
quantitative study involving the observed exchange
between the alkyl groups in the ion pairs exo-2b-B1
and exo-2b-B2 via 1H-1H EXSY experiments (Figure
8, Table 2, entry 7). In these ion pairs, the signals for
Scheme 9
Hayes et al.
the two backbone protons were not baseline resolved at
400 MHz, and so the signals for the SiMe2, SiMe3, and
SiCH2Si protons were used to assess the rate of ex-
change; this allowed for an internal check on the
precision of the data, and the values obtained at 244 K
in Table 2 reflect rate constants obtained from the
different resonance pairs as labeled in Figure 8. From
EXSY experiments run at various temperatures, an
Eyring plot was obtained (see Figure S4), and the
thermodynamic parameters given in Table 2, entry 7,
were obtained. The activation parameters (∆Hq ) 8.7(3)
kcal mol-1, ∆Sq ) -20(2) eu) differ substantially from
those obtained for the intermolecular ion exchanges
observed between the scandium methyl ion pairs. In
particular, the entropy of activation is significantly more
negative, while the activation enthalpy is lower by 2-3
kcal mol-1.
Taken together, these results may be interpreted in
terms of a two-step mechanism that involves reversible
(and rate limiting) associative solvation of the CIPs to
form solvent-separated ion pair (SSIP) intermediates
that can undergo rapid exchange of ions before collaps-
ing back to a CIP mixture in which the ions have been
statistically randomized (Scheme 9). The negative ∆Sq
values are consistent with associative displacement of
the methylborate anions by solvent, in this case toluene.
In the ion pairs exo-2b-Bn, the values are more
negative and reflective of the intrinsic entropy of activa-
tion for this process, while in the endo/exo-2a-Bn ion
pairs, ligand flipping attenuates the observed value
because of the higher flexibility of this system. The ∆Hq
values are lower in the 2b ion pairs since the much
bulkier alkyl group encourages displacement of the
anion from the inner coordination sphere to an outer
sphere domain.
The nature of the SSIPs in Scheme 9 is supported by
the isolation and characterization of a family of such
species supported by the LMe β-diketiminato ligand (e.g.,
Ion Pair Dynamics
III) that we have previously reported.21 In these com-
pounds, the arene solvents are observed to coordinate
to the scandium center exclusively in the exo position
with an unsymmetrical η6-bonding mode and are not
readily displaced by the more weakly coordinating
[B(C6F5)4]- anion employed. However, the coordinated
arenes are in exchange with free solvent molecules, and
an analysis of the activation parameters for this ex-
change suggests that slippage from the ground state η6
structure to an intermediate of lower hapticity is the
first step in the process.21 In compounds 2 discussed
herein, the less open coordination environment fostered
by the LtBu β-diketiminato ancillary and the more
donating nature of the [MeBn]- anion results in higher
energies for these η6-toluene SSIPs, but they remain
viable intermediates in the present chemistry. It is
possible that η1-4 bonding modes similar to those
proposed in computational investigations of toluene-
solvated metallocenium ions45 may be operative in the
SSIPs in this system. A toluene-solvated ion pair of the
“constrained geometry” zirconium ion pair [Me2Si(C5-
Me4)(t-BuN)ZrMe(toluene)]+[B(C6F5)4]- has been char-
acterized by NMR spectroscopy.46
That these species mediate the intramolecular ion
swapping observed is supported by the recent detailed
studies on the solution structures and ion pairing
behavior of a variety of metallocenium CIP and SSIP
derivatives by Marks et al.12b While not explicitly
observing ion promiscuity in the same sense that we
have with the crossover studies reported above, this
behavior is implicit in the observed tendency of SSIPs
to aggregate at moderate concentrations, while CIPs
remain discrete ion pairs under all conditions studied
with specific and well-defined anion-cation interactions.
In the SSIPs studied, the anions were observed to be
paired with the cations but were much more flexible in
the position assumed about the cation environment.
Further evidence in support of the involvement of SSIPs
in ion mobility is found in an observation reported by
Brintzinger, where rapid anion exchange between an-
(45) (a) Vanka, K.; Chan, M. S. W.; Pye, C. C.; Ziegler, T. Organo-
metallics 2000, 19, 1841. (b) Chan, M. S. W.; Vanka, K.; Pye, C. C.;
Ziegler, T. Organometallics 1999, 18, 4624.
(46) Jia, L.; Yang, X.; Stern, C. L.; Marks, T. J. Organometallics
1997, 16, 842.
Organometallics, Vol. 24, No. 6, 2005 1183
ions in the CIP [Me2Si(C5H4)2ZrMe]+[MeB(C6F5)3]- and
the SSIP model [Me2Si(C5H4)2ZrMe(NMe2Ph)]+[MeB-
(C6F5)3]- is “turned on” only when appreciable quanti-
ties of the SSIP are generated upon treatment of the
CIP with NMe2Ph.8c Whether or not formal aggregates
are intimately involved in the exchange of ions between
ion pairs is not clear but seems reasonable to invoke in
the proper concentration regime.
Conclusions
In this paper, we have reported the synthesis and
detailed solution behavior of a family of well-defined
organoscandium cations with relevance to species active
in olefin polymerization. To our knowledge, this is the
first such study of ion pairs that are not based on the
metallocenium ligand framework (or the closely related
“constrained geometry” systems). The data we have
acquired concerning the ion mobility within and be-
tween ion pairs add to the library of available studies
concerning these important processes and support a
general picture wherein intermolecular ion exchange is
facile in SSIPs but not in CIPs. In ion pairs involving
the relatively sticky [MeB(C6F6)3]- anionic partner, the
CIP is the ground state, and rate-limiting generation
of a higher energy SSIP (via an associative displacement
of the anion by solvent) is required to initiate rapid ion
interchange. In the present study, a direct comparison
between scandium methyl ion pairs 1 and ion pairs 2,
which incorporate a bulkier alkyl ligand more similar
to a growing polymeryl group, reveals that ion pair
reorganization is a lower energy process in compounds
2 than the methyl ion pairs 1. This corroborates similar
studies with metallocenium ions,11 but the lower en-
thalpic barrier to anion displacement from the β-diketim-
inato scandium cations incorporating the bulky alkyl
group is somewhat tempered by a higher entropic
barrier due to a restriction on the fluxionality of the
ancillary ligand in these more sterically crowded sys-
tems.
Acknowledgment. Financial support for this work
was provided by NSERC of Canada in the form of a
Discovery Grant to W.E.P. and scholarship support to
P.G.H. (PGS-A and PGS-B). P.G.H. also thanks the
Alberta Heritage Foundation for a Steinhauer Award
and the Sir Izaak Walton Killam Foundation for a
Doctoral Fellowship.
Supporting Information Available: Full experimental
details, compound characterization data, representative spec-
tra, first-order plots, and Eyring plot for the metalation of
exo-2b-B1 and crystallographic information files (CIF) for
[LMeScMe]+[MeB(C6F5)3]- and exo-2b-B1. This material is
available free of charge via the Internet at http://pubs.acs.org.
OM050007V